STK16C88-3

Device Operation

The AutoStore+ STK16C88-3 is a fast 32K x 8 SRAM that does not lose its data on power down. The data is preserved in integral QuantumTrap non-volatile storage elements when power is lost. Automatic STORE on power down and automatic RECALL on power up guarantee data integrity without the use of batteries.

SRAM Read

The STK16C88-3 performs a READ cycle whenever CE and OE are LOW while WE is HIGH. The address specified on pins A0–14determines the 32,768 data bytes accessed. When the READ is initiated by an address transition, the outputs are valid after a delay of tAA (READ cycle 1). If the READ is initiated by CE or OE, the outputs are valid at tACE or at tDOE, whichever is later (READ cycle 2). The data outputs repeatedly respond to address changes within the tAA access time without the need for transi- tions on any control input pins, and remains valid until another address change or until CE or OE is brought HIGH.

SRAM Write

A WRITE cycle is performed whenever CE and WE are LOW. The address inputs must be stable prior to entering the WRITE cycle and must remain stable until either CE or WE goes HIGH at the end of the cycle. The data on the common IO pins DQ0–7are written into the memory if it has valid tSD, before the end of a WE controlled WRITE or before the end of an CE controlled WRITE. Keep OE HIGH during the entire WRITE cycle to avoid data bus contention on common IO lines. If OE is left LOW, internal circuitry turns off the output buffers tHZWE after WE goes LOW.

AutoStore+ Operation

The STK16C88-3’s automatic STORE on power down is com- pletely transparent to the system. The STORE initiation takes less than 500 ns when power is lost (VCC<VSWITCH) at which point the part depends only on its internal capacitor for STORE com- pletion.

If the power supply drops faster than 20 μs/volt before Vcc reaches Vswitch, then a 2.2 ohm resistor should be inserted between Vcc and the system supply to avoid a momentary excess of current between Vcc and internal capacitor.

In order to prevent unneeded STORE operations, automatic STOREs are ignored unless at least one WRITE operation has taken place since the most recent STORE or RECALL cycle. Software initiated STORE cycles are performed regardless of whether or not a WRITE operation has taken place.

Hardware RECALL (Power Up)

If the STK16C88-3 is in a WRITE state at the end of power up RECALL, the SRAM data is corrupted. To help avoid this situation, a 10 Kohm resistor is connected either between WE and system VCC or between CE and system VCC.

Software STORE

Data is transferred from the SRAM to the nonvolatile memory by a software address sequence. The STK16C88-3 software STORE cycle is initiated by executing sequential CE controlled READ cycles from six specific address locations in exact order. During the STORE cycle, an erase of the previous nonvolatile data is first performed followed by a program of the nonvolatile elements. When a STORE cycle is initiated, input and output are disabled until the cycle is completed.

Because a sequence of READs from specific addresses is used for STORE initiation, it is important that no other READ or WRITE accesses intervene in the sequence. If they intervene, the sequence is aborted and no STORE or RECALL takes place.

To initiate the software STORE cycle, the following READ sequence is performed:

1.Read address 0x0E38, Valid READ

2.Read address 0x31C7, Valid READ

3.Read address 0x03E0, Valid READ

4.Read address 0x3C1F, Valid READ

5.Read address 0x303F, Valid READ

6.Read address 0x0FC0, Initiate STORE cycle

The software sequence is clocked with CE controlled READs. When the sixth address in the sequence is entered, the STORE cycle commences and the chip is disabled. It is important that READ cycles and not WRITE cycles are used in the sequence. It is not necessary that OE is LOW for a valid sequence. After the tSTORE cycle time is fulfilled, the SRAM is again activated for READ and WRITE operation.

Software RECALL

Data is transferred from the nonvolatile memory to the SRAM by a software address sequence. A software RECALL cycle is initiated with a sequence of READ operations in a manner similar to the software STORE initiation. To initiate the RECALL cycle, the following sequence of CE controlled READ operations is performed:

1.Read address 0x0E38, Valid READ

2.Read address 0x31C7, Valid READ

3.Read address 0x03E0, Valid READ

4.Read address 0x3C1F, Valid READ

5.Read address 0x303F, Valid READ

6.Read address 0x0C63, Initiate RECALL cycle

During power up or after any low power condition (VCC<VRESET), an internal RECALL request is latched. When VCC once again exceeds the sense voltage of VSWITCH, a RECALL cycle is automatically initiated and takes tHRECALL to complete.

Internally, RECALL is a two step procedure. First, the SRAM data is cleared, and then the nonvolatile information is transferred into the SRAM cells. After the tRECALL cycle time, the SRAM is once again ready for READ and WRITE operations. The RECALL operation does not alter the data in the nonvolatile elements. The nonvolatile data can be recalled an unlimited number of times.

Document Number: 001-50594 Rev. **

Page 3 of 14

[+] Feedback

Page 3
Image 3
Cypress STK16C88-3 Device Operation, Sram Read, Sram Write, AutoStore+ Operation, Hardware Recall Power Up, Software Store